Decision Integrity Check
A reviewable output structure for material decisions.
The Decision Integrity Check is the proposed structured record produced through the Planetary Decision Integrity Protocol and supported by Kairo, the planned Consequence Intelligence engine.
Its purpose is not to approve or reject decisions automatically. Its purpose is to make the reasoning around major decisions visible enough to be questioned, reviewed, improved and governed.
01
Purpose
Major decisions often move through institutions with fragmented evidence, implicit assumptions, incomplete visibility of affected systems and limited traceability of trade-offs.
The Decision Integrity Check is intended to create a minimum reviewable record before such decisions become reality.
It asks a simple question:
The check is designed to support decision integrity by making visible:
What must be visible before a material decision can be considered responsibly understood?
- the decision being considered;
- the authority and actors involved;
- the systems likely to be affected;
- the evidence used;
- the assumptions and exclusions made;
- the uncertainty that remains;
- the alternatives considered;
- the trade-offs accepted;
- the human review applied.
02
What the check records
The Decision Integrity Check is not a report written after the fact. It is intended as a structured decision record prepared before action is taken.
It records six essential dimensions.
- 01
Decision scope
What is being decided, by whom, under which authority, over what timeframe, and with what material consequences.
- 02
Affected systems
Which planetary, territorial, institutional, social, economic or technological systems may be affected by the decision.
- 03
Evidence register
Which sources, models, datasets, documents, expert inputs and prior assessments support the consequence analysis.
- 04
Assumption register
Which assumptions, limits, exclusions, dependencies and methodological choices shape the reasoning.
- 05
Uncertainty and alternatives
What remains uncertain, what alternative paths exist, and which trade-offs are being accepted or deferred.
- 06
Human review
Who reviews the assessment, which authority remains human, what objections were considered, and how accountability is preserved.
03
Core output structure
A Decision Integrity Check should be structured enough to be comparable across decisions, but flexible enough to adapt to different domains.
The proposed minimum structure is:
- 01
Decision summary
A concise description of the material decision, its scope, decision authority, timing and expected action pathway.
- 02
Consequence map
A structured map of direct, indirect and systemic effects across relevant systems and time horizons.
- 03
Evidence register
A traceable list of the sources and inputs used to support the assessment.
- 04
Assumption register
A transparent record of assumptions, exclusions, limits and dependencies.
- 05
Uncertainty register
A clear distinction between what is known, what is estimated, what is contested and what remains unknown.
- 06
Alternatives and trade-offs
A review of plausible alternatives, rejected options, accepted trade-offs and possible mitigation pathways.
- 07
Human review record
A record of review authority, reviewers, objections, unresolved concerns and final responsibility.
04
Example output skeleton
The following skeleton is illustrative. It is not presented as a final protocol template.
Illustrative skeleton
Decision Integrity Check
Version: draft / founding phase
Decision ID: [to be assigned]
Decision owner: [institution / authority]
Decision type: [policy / investment / infrastructure / technology / resource allocation]
Review status: [draft / reviewed / contested / approved for decision review]
1. Decision Summary
- What is being decided?
- Who has authority?
- What action follows the decision?
- What is the expected timeframe?
2. Affected Systems
- Planetary systems
- Territorial systems
- Institutional systems
- Social and economic systems
- Technological systems
3. Evidence Register
- Sources used
- Models or datasets used
- Expert inputs
- Documents reviewed
- Evidence limits
4. Assumption Register
- Key assumptions
- Exclusions
- Dependencies
- Methodological limits
5. Consequence Assessment
- Direct effects
- Indirect effects
- Systemic effects
- Time horizons
- Reversibility / persistence
6. Uncertainty and Trade-offs
- Known uncertainties
- Contested points
- Alternative pathways
- Accepted trade-offs
- Deferred consequences
7. Human Review
- Reviewers
- Objections raised
- Unresolved concerns
- Accountability owner
- Final review note 05
Human review
The Decision Integrity Check is designed around a non-negotiable principle:
Kairo may help connect evidence, structure reasoning, expose uncertainty and generate reviewable outputs. Human authority remains responsible for judgment, legitimacy, interpretation and final institutional action.
A valid Decision Integrity Check must therefore preserve:
AI may support consequence reasoning. It must not become the authority that legitimizes the decision.
- explicit human review;
- traceable reasoning;
- visible uncertainty;
- reviewable assumptions;
- documented objections;
- clear accountability.
06
Boundaries
The Decision Integrity Check does not replace legal duties, scientific review, democratic legitimacy, political accountability, professional judgment or institutional responsibility.
It does not certify that a decision is good.
It does not claim that every consequence can be predicted.
It does not transfer responsibility to an AI system.
It creates a structured record so that consequence reasoning can be examined before power turns into reality.
07
Current status
The Decision Integrity Check is under definition during the founding phase of Planetary Integrity Foundation.
The current priority is to define:
Continue with the Protocol or the Founding Brief to understand the broader framework.
- the minimum check structure;
- the relationship with the Planetary Decision Integrity Protocol;
- the role of Kairo in generating reviewable outputs;
- the evidence and assumption standards;
- the human review model;
- the contributor process for technical, legal, scientific and institutional review.